JS speed - Konq vs. Mozilla

[Koos Vriezen]

On Thu, 5 Sep 2002, Waldo Bastian wrote:

> On Thursday 05 September 2002 10:31 am, Carsten Pfeiffer wrote:
> > On Thursday 05 September 2002 18:11, Koos Vriezen wrote:
> > > On Thu, 5 Sep 2002, Carsten Pfeiffer wrote:
> > > > Isn't the '&' part of the arguments-signature so that you could
> > > > overload putValue()? Not sure which version would be used then, tho.
> > >
> > > You would get a 'call of overloaded `putValue(..)' is ambiguous. Even if
> >
> > Thanks, that makes sense.
> >
> > > we could avoid it, an extra virtual putValue would break BC.
>
> you can add an extra putValueNewAPI() function together with a strategically
> placed #define putValue putValueNewAPI.

I don't follow this. You get a libkjs with putValueNewAPI (different
signature), with putValue (no change) or with both (larger vtable). How
can that be BC?
The only thing I can think of is building a libkjsNewAPI next to libkjs.
So, what the secret that khtml/ecma sees another libkjs than an app linked
against the old libkjs?

Koos